Digital Audio and Signal Processing

Digital audio represents sound as a sequence of numbers rather than a continuously varying voltage. A continuous sound wave is captured by sampling its amplitude at regular intervals and quantising each sample to a numeric value, a process performed by an analog-to-digital converter. Two parameters dominate the result: the sampling rate, which sets the highest frequency that can be represented, and the bit depth, which determines the number of amplitude levels and therefore the dynamic range and quantisation noise. Played back, the numbers are converted to an analog signal by a digital-to-analog converter.

Once in numeric form, audio can be manipulated by digital signal processing (DSP), in which arithmetic operations carry out tasks that would be difficult in analog hardware. Digital filters, designed to pass or attenuate chosen frequency ranges, are central tools for equalisation, noise reduction and effects. Working in the digital domain also raises specific concerns, including the signal-to-noise ratio set by bit depth and timing errors known as jitter, which can degrade quality if the sampling clock is unstable.

Because uncompressed audio occupies a large amount of data, compression is widely used. Lossless methods reduce size without discarding information, while lossy methods such as MP3 exploit limitations of human hearing to discard inaudible detail and achieve much smaller files. Multichannel and broadcast formats include AC-3 (Dolby Digital) and DTS, used for surround sound on optical discs and broadcasting, while DSD is a high-resolution one-bit representation. Each scheme balances file size, quality and processing complexity for its intended use.

Frequently asked questions

What do sampling rate and bit depth control?

The sampling rate sets the highest frequency that can be captured, while the bit depth sets the number of amplitude levels, which determines dynamic range and quantisation noise.

What is jitter in digital audio?

Jitter is timing instability in the sampling or playback clock. If the intervals between samples vary, the reconstructed waveform is distorted, degrading sound quality.

How does lossy compression like MP3 reduce file size?

Lossy codecs discard audio detail that human hearing is unlikely to notice, using perceptual models, which shrinks the data far more than lossless methods at the cost of some fidelity.